Automatic generation of skeletal mechanisms for ignition combustion based on level of importance analysis
- School of Engineering and Materials Sciences, Queen Mary University of London, London E1 4NS (United Kingdom)
A level of importance (LOI) selection parameter is employed in order to identify species with general low importance to the overall accuracy of a chemical model. This enables elimination of the minor reaction paths in which these species are involved. The generation of such skeletal mechanisms is performed automatically in a pre-processing step ranking species according to their level of importance. This selection criterion is a combined parameter based on a time scale and sensitivity analysis, identifying both short lived species and species with respect to which the observable of interest has low sensitivity. In this work a careful element flux analysis demonstrates that such species do not interact in major reaction paths. Employing the LOI procedure replaces the previous method of identifying redundant species through a two step procedure involving a reaction flow analysis followed by a sensitivity analysis. The flux analysis is performed using DARS {sup copyright}, a digital analysis tool modelling reactive systems. Simplified chemical models are generated based on a detailed ethylene mechanism involving 111 species and 784 reactions (1566 forward and backward reactions) proposed by Wang et al. Eliminating species from detailed mechanisms introduces errors in the predicted combustion parameters. In the present work these errors are systematically studied for a wide range of conditions, including temperature, pressure and mixtures. Results show that the accuracy of simplified models is particularly lowered when the initial temperatures are close to the transition between low- and high-temperature chemistry. A speed-up factor of 5 is observed when using a simplified model containing only 27% of the original species and 19% of the original reactions. (author)
- OSTI ID:
- 21195822
- Journal Information:
- Combustion and Flame, Vol. 156, Issue 7; Other Information: Elsevier Ltd. All rights reserved; ISSN 0010-2180
- Country of Publication:
- United States
- Language:
- English
Similar Records
An efficient error-propagation-based reduction method for large chemical kinetic mechanisms
On the application of betweenness centrality in chemical network analysis: Computational diagnostics and model reduction
Related Subjects
ORGANIC
PHYSICAL AND ANALYTICAL CHEMISTRY
COMBUSTION
AUTOIGNITION
TEMPERATURE RANGE 1000-4000 K
SENSITIVITY ANALYSIS
ACCURACY
ERRORS
MIXTURES
COMBUSTION KINETICS
SIMULATION
TEMPERATURE DEPENDENCE
PRESSURE DEPENDENCE
PRESSURE RANGE KILO PA
MATHEMATICAL MODELS
Chemical kinetics reduction
Numerical methods
Level of importance method